http.c revision 87561 
1/*-
2 * Copyright (c) 2000 Dag-Erling Co�dan Sm�rgrav
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer
10 *    in this position and unchanged.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 *    notice, this list of conditions and the following disclaimer in the
13 *    documentation and/or other materials provided with the distribution.
14 * 3. The name of the author may not be used to endorse or promote products
15 *    derived from this software without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27 */
28
29#include <sys/cdefs.h>
30__FBSDID("$FreeBSD: head/lib/libfetch/http.c 87561 2001-12-09 15:05:58Z des $");
31
32/*
33 * The following copyright applies to the base64 code:
34 *
35 *-
36 * Copyright 1997 Massachusetts Institute of Technology
37 *
38 * Permission to use, copy, modify, and distribute this software and
39 * its documentation for any purpose and without fee is hereby
40 * granted, provided that both the above copyright notice and this
41 * permission notice appear in all copies, that both the above
42 * copyright notice and this permission notice appear in all
43 * supporting documentation, and that the name of M.I.T. not be used
44 * in advertising or publicity pertaining to distribution of the
45 * software without specific, written prior permission.  M.I.T. makes
46 * no representations about the suitability of this software for any
47 * purpose.  It is provided "as is" without express or implied
48 * warranty.
49 *
50 * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''.  M.I.T. DISCLAIMS
51 * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE,
52 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
53 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT
54 * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
55 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
56 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
57 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
58 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
59 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
60 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
61 * SUCH DAMAGE.
62 */
63
64#include <sys/param.h>
65#include <sys/socket.h>
66
67#include <ctype.h>
68#include <err.h>
69#include <errno.h>
70#include <locale.h>
71#include <netdb.h>
72#include <stdarg.h>
73#include <stdio.h>
74#include <stdlib.h>
75#include <string.h>
76#include <time.h>
77#include <unistd.h>
78
79#include "fetch.h"
80#include "common.h"
81#include "httperr.h"
82
83extern char *__progname; /* XXX not portable */
84
85/* Maximum number of redirects to follow */
86#define MAX_REDIRECT 5
87
88/* Symbolic names for reply codes we care about */
89#define HTTP_OK			200
90#define HTTP_PARTIAL		206
91#define HTTP_MOVED_PERM		301
92#define HTTP_MOVED_TEMP		302
93#define HTTP_SEE_OTHER		303
94#define HTTP_NEED_AUTH		401
95#define HTTP_NEED_PROXY_AUTH	407
96#define HTTP_PROTOCOL_ERROR	999
97
98#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \
99                            || (xyz) == HTTP_MOVED_TEMP \
100                            || (xyz) == HTTP_SEE_OTHER)
101
102
103
104/*****************************************************************************
105 * I/O functions for decoding chunked streams
106 */
107
108struct cookie
109{
110    int		 fd;
111    char	*buf;
112    size_t	 b_size;
113    ssize_t	 b_len;
114    int		 b_pos;
115    int		 eof;
116    int		 error;
117    size_t	 chunksize;
118#ifndef NDEBUG
119    size_t	 total;
120#endif
121};
122
123/*
124 * Get next chunk header
125 */
126static int
127_http_new_chunk(struct cookie *c)
128{
129    char *p;
130
131    if (_fetch_getln(c->fd, &c->buf, &c->b_size, &c->b_len) == -1)
132	return -1;
133
134    if (c->b_len < 2 || !ishexnumber(*c->buf))
135	return -1;
136
137    for (p = c->buf; !isspace(*p) && *p != ';' && p < c->buf + c->b_len; ++p)
138	if (!ishexnumber(*p))
139	    return -1;
140	else if (isdigit(*p))
141	    c->chunksize = c->chunksize * 16 + *p - '0';
142	else
143	    c->chunksize = c->chunksize * 16 + 10 + tolower(*p) - 'a';
144
145#ifndef NDEBUG
146    if (fetchDebug) {
147	c->total += c->chunksize;
148	if (c->chunksize == 0)
149	    fprintf(stderr, "\033[1m_http_fillbuf(): "
150		    "end of last chunk\033[m\n");
151	else
152	    fprintf(stderr, "\033[1m_http_fillbuf(): "
153		    "new chunk: %lu (%lu)\033[m\n",
154		    (unsigned long)c->chunksize, (unsigned long)c->total);
155    }
156#endif
157
158    return c->chunksize;
159}
160
161/*
162 * Fill the input buffer, do chunk decoding on the fly
163 */
164static int
165_http_fillbuf(struct cookie *c)
166{
167    if (c->error)
168	return -1;
169    if (c->eof)
170	return 0;
171
172    if (c->chunksize == 0) {
173	switch (_http_new_chunk(c)) {
174	case -1:
175	    c->error = 1;
176	    return -1;
177	case 0:
178	    c->eof = 1;
179	    return 0;
180	}
181    }
182
183    if (c->b_size < c->chunksize) {
184	char *tmp;
185
186	if ((tmp = realloc(c->buf, c->chunksize)) == NULL)
187	    return -1;
188	c->buf = tmp;
189	c->b_size = c->chunksize;
190    }
191
192    if ((c->b_len = read(c->fd, c->buf, c->chunksize)) == -1)
193	return -1;
194    c->chunksize -= c->b_len;
195
196    if (c->chunksize == 0) {
197	char endl[2];
198	read(c->fd, endl, 2);
199    }
200
201    c->b_pos = 0;
202
203    return c->b_len;
204}
205
206/*
207 * Read function
208 */
209static int
210_http_readfn(void *v, char *buf, int len)
211{
212    struct cookie *c = (struct cookie *)v;
213    int l, pos;
214
215    if (c->error)
216	return -1;
217    if (c->eof)
218	return 0;
219
220    for (pos = 0; len > 0; pos += l, len -= l) {
221	/* empty buffer */
222	if (!c->buf || c->b_pos == c->b_len)
223	    if (_http_fillbuf(c) < 1)
224		break;
225	l = c->b_len - c->b_pos;
226	if (len < l)
227	    l = len;
228	bcopy(c->buf + c->b_pos, buf + pos, l);
229	c->b_pos += l;
230    }
231
232    if (!pos && c->error)
233	return -1;
234    return pos;
235}
236
237/*
238 * Write function
239 */
240static int
241_http_writefn(void *v, const char *buf, int len)
242{
243    struct cookie *c = (struct cookie *)v;
244
245    return write(c->fd, buf, len);
246}
247
248/*
249 * Close function
250 */
251static int
252_http_closefn(void *v)
253{
254    struct cookie *c = (struct cookie *)v;
255    int r;
256
257    r = close(c->fd);
258    if (c->buf)
259	free(c->buf);
260    free(c);
261    return r;
262}
263
264/*
265 * Wrap a file descriptor up
266 */
267static FILE *
268_http_funopen(int fd)
269{
270    struct cookie *c;
271    FILE *f;
272
273    if ((c = calloc(1, sizeof *c)) == NULL) {
274	_fetch_syserr();
275	return NULL;
276    }
277    c->fd = fd;
278    if (!(f = funopen(c, _http_readfn, _http_writefn, NULL, _http_closefn))) {
279	_fetch_syserr();
280	free(c);
281	return NULL;
282    }
283    return f;
284}
285
286
287/*****************************************************************************
288 * Helper functions for talking to the server and parsing its replies
289 */
290
291/* Header types */
292typedef enum {
293    hdr_syserror = -2,
294    hdr_error = -1,
295    hdr_end = 0,
296    hdr_unknown = 1,
297    hdr_content_length,
298    hdr_content_range,
299    hdr_last_modified,
300    hdr_location,
301    hdr_transfer_encoding,
302    hdr_www_authenticate
303} hdr_t;
304
305/* Names of interesting headers */
306static struct {
307    hdr_t	 num;
308    const char	*name;
309} hdr_names[] = {
310    { hdr_content_length,	"Content-Length" },
311    { hdr_content_range,	"Content-Range" },
312    { hdr_last_modified,	"Last-Modified" },
313    { hdr_location,		"Location" },
314    { hdr_transfer_encoding,	"Transfer-Encoding" },
315    { hdr_www_authenticate,	"WWW-Authenticate" },
316    { hdr_unknown,		NULL },
317};
318
319static char	*reply_buf;
320static size_t	 reply_size;
321static size_t	 reply_length;
322
323/*
324 * Send a formatted line; optionally echo to terminal
325 */
326static int
327_http_cmd(int fd, const char *fmt, ...)
328{
329    va_list ap;
330    size_t len;
331    char *msg;
332    int r;
333
334    va_start(ap, fmt);
335    len = vasprintf(&msg, fmt, ap);
336    va_end(ap);
337
338    if (msg == NULL) {
339	errno = ENOMEM;
340	_fetch_syserr();
341	return -1;
342    }
343
344    r = _fetch_putln(fd, msg, len);
345    free(msg);
346
347    if (r == -1) {
348	_fetch_syserr();
349	return -1;
350    }
351
352    return 0;
353}
354
355/*
356 * Get and parse status line
357 */
358static int
359_http_get_reply(int fd)
360{
361    char *p;
362
363    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
364	return -1;
365    /*
366     * A valid status line looks like "HTTP/m.n xyz reason" where m
367     * and n are the major and minor protocol version numbers and xyz
368     * is the reply code.
369     * Unfortunately, there are servers out there (NCSA 1.5.1, to name
370     * just one) that do not send a version number, so we can't rely
371     * on finding one, but if we do, insist on it being 1.0 or 1.1.
372     * We don't care about the reason phrase.
373     */
374    if (strncmp(reply_buf, "HTTP", 4) != 0)
375	return HTTP_PROTOCOL_ERROR;
376    p = reply_buf + 4;
377    if (*p == '/') {
378	if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1'))
379	    return HTTP_PROTOCOL_ERROR;
380	p += 4;
381    }
382    if (*p != ' '
383	|| !isdigit(p[1])
384	|| !isdigit(p[2])
385	|| !isdigit(p[3]))
386	return HTTP_PROTOCOL_ERROR;
387
388    return ((p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'));
389}
390
391/*
392 * Check a header; if the type matches the given string, return a
393 * pointer to the beginning of the value.
394 */
395static const char *
396_http_match(const char *str, const char *hdr)
397{
398    while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
399	/* nothing */;
400    if (*str || *hdr != ':')
401	return NULL;
402    while (*hdr && isspace(*++hdr))
403	/* nothing */;
404    return hdr;
405}
406
407/*
408 * Get the next header and return the appropriate symbolic code.
409 */
410static hdr_t
411_http_next_header(int fd, const char **p)
412{
413    int i;
414
415    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
416	return hdr_syserror;
417    while (reply_length && isspace(reply_buf[reply_length-1]))
418	reply_length--;
419    reply_buf[reply_length] = 0;
420    if (reply_length == 0)
421	return hdr_end;
422    /*
423     * We could check for malformed headers but we don't really care.
424     * A valid header starts with a token immediately followed by a
425     * colon; a token is any sequence of non-control, non-whitespace
426     * characters except "()<>@,;:\\\"{}".
427     */
428    for (i = 0; hdr_names[i].num != hdr_unknown; i++)
429	if ((*p = _http_match(hdr_names[i].name, reply_buf)) != NULL)
430	    return hdr_names[i].num;
431    return hdr_unknown;
432}
433
434/*
435 * Parse a last-modified header
436 */
437static int
438_http_parse_mtime(const char *p, time_t *mtime)
439{
440    char locale[64], *r;
441    struct tm tm;
442
443    strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
444    setlocale(LC_TIME, "C");
445    r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
446    /* XXX should add support for date-2 and date-3 */
447    setlocale(LC_TIME, locale);
448    if (r == NULL)
449	return -1;
450    DEBUG(fprintf(stderr, "last modified: [\033[1m%04d-%02d-%02d "
451		  "%02d:%02d:%02d\033[m]\n",
452		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
453		  tm.tm_hour, tm.tm_min, tm.tm_sec));
454    *mtime = timegm(&tm);
455    return 0;
456}
457
458/*
459 * Parse a content-length header
460 */
461static int
462_http_parse_length(const char *p, off_t *length)
463{
464    off_t len;
465
466    for (len = 0; *p && isdigit(*p); ++p)
467	len = len * 10 + (*p - '0');
468    if (*p)
469	return -1;
470    DEBUG(fprintf(stderr, "content length: [\033[1m%lld\033[m]\n",
471		  (long long)len));
472    *length = len;
473    return 0;
474}
475
476/*
477 * Parse a content-range header
478 */
479static int
480_http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size)
481{
482    off_t first, last, len;
483
484    if (strncasecmp(p, "bytes ", 6) != 0)
485	return -1;
486    for (first = 0, p += 6; *p && isdigit(*p); ++p)
487	first = first * 10 + *p - '0';
488    if (*p != '-')
489	return -1;
490    for (last = 0, ++p; *p && isdigit(*p); ++p)
491	last = last * 10 + *p - '0';
492    if (first > last || *p != '/')
493	return -1;
494    for (len = 0, ++p; *p && isdigit(*p); ++p)
495	len = len * 10 + *p - '0';
496    if (*p || len < last - first + 1)
497	return -1;
498    DEBUG(fprintf(stderr, "content range: [\033[1m%lld-%lld/%lld\033[m]\n",
499		  (long long)first, (long long)last, (long long)len));
500    *offset = first;
501    *length = last - first + 1;
502    *size = len;
503    return 0;
504}
505
506
507/*****************************************************************************
508 * Helper functions for authorization
509 */
510
511/*
512 * Base64 encoding
513 */
514static char *
515_http_base64(char *src)
516{
517    static const char base64[] =
518	"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
519	"abcdefghijklmnopqrstuvwxyz"
520	"0123456789+/";
521    char *str, *dst;
522    size_t l;
523    int t, r;
524
525    l = strlen(src);
526    if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
527	return NULL;
528    dst = str;
529    r = 0;
530
531    while (l >= 3) {
532	t = (src[0] << 16) | (src[1] << 8) | src[2];
533	dst[0] = base64[(t >> 18) & 0x3f];
534	dst[1] = base64[(t >> 12) & 0x3f];
535	dst[2] = base64[(t >> 6) & 0x3f];
536	dst[3] = base64[(t >> 0) & 0x3f];
537	src += 3; l -= 3;
538	dst += 4; r += 4;
539    }
540
541    switch (l) {
542    case 2:
543	t = (src[0] << 16) | (src[1] << 8);
544	dst[0] = base64[(t >> 18) & 0x3f];
545	dst[1] = base64[(t >> 12) & 0x3f];
546	dst[2] = base64[(t >> 6) & 0x3f];
547	dst[3] = '=';
548	dst += 4;
549	r += 4;
550	break;
551    case 1:
552	t = src[0] << 16;
553	dst[0] = base64[(t >> 18) & 0x3f];
554	dst[1] = base64[(t >> 12) & 0x3f];
555	dst[2] = dst[3] = '=';
556	dst += 4;
557	r += 4;
558	break;
559    case 0:
560	break;
561    }
562
563    *dst = 0;
564    return str;
565}
566
567/*
568 * Encode username and password
569 */
570static int
571_http_basic_auth(int fd, const char *hdr, const char *usr, const char *pwd)
572{
573    char *upw, *auth;
574    int r;
575
576    DEBUG(fprintf(stderr, "usr: [\033[1m%s\033[m]\n", usr));
577    DEBUG(fprintf(stderr, "pwd: [\033[1m%s\033[m]\n", pwd));
578    if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
579	return -1;
580    auth = _http_base64(upw);
581    free(upw);
582    if (auth == NULL)
583	return -1;
584    r = _http_cmd(fd, "%s: Basic %s", hdr, auth);
585    free(auth);
586    return r;
587}
588
589/*
590 * Send an authorization header
591 */
592static int
593_http_authorize(int fd, const char *hdr, const char *p)
594{
595    /* basic authorization */
596    if (strncasecmp(p, "basic:", 6) == 0) {
597	char *user, *pwd, *str;
598	int r;
599
600	/* skip realm */
601	for (p += 6; *p && *p != ':'; ++p)
602	    /* nothing */ ;
603	if (!*p || strchr(++p, ':') == NULL)
604	    return -1;
605	if ((str = strdup(p)) == NULL)
606	    return -1; /* XXX */
607	user = str;
608	pwd = strchr(str, ':');
609	*pwd++ = '\0';
610	r = _http_basic_auth(fd, hdr, user, pwd);
611	free(str);
612	return r;
613    }
614    return -1;
615}
616
617
618/*****************************************************************************
619 * Helper functions for connecting to a server or proxy
620 */
621
622/*
623 * Connect to the correct HTTP server or proxy.
624 */
625static int
626_http_connect(struct url *URL, struct url *purl, const char *flags)
627{
628    int verbose;
629    int af, fd;
630
631#ifdef INET6
632    af = AF_UNSPEC;
633#else
634    af = AF_INET;
635#endif
636
637    verbose = CHECK_FLAG('v');
638    if (CHECK_FLAG('4'))
639	af = AF_INET;
640#ifdef INET6
641    else if (CHECK_FLAG('6'))
642	af = AF_INET6;
643#endif
644
645    if (purl) {
646	URL = purl;
647    } else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) {
648	/* can't talk http to an ftp server */
649	/* XXX should set an error code */
650	return -1;
651    }
652
653    if ((fd = _fetch_connect(URL->host, URL->port, af, verbose)) == -1)
654	/* _fetch_connect() has already set an error code */
655	return -1;
656    return fd;
657}
658
659static struct url *
660_http_get_proxy(void)
661{
662    struct url *purl;
663    char *p;
664
665    if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) &&
666	(purl = fetchParseURL(p))) {
667	if (!*purl->scheme)
668	    strcpy(purl->scheme, SCHEME_HTTP);
669	if (!purl->port)
670	    purl->port = _fetch_default_proxy_port(purl->scheme);
671	if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0)
672	    return purl;
673	fetchFreeURL(purl);
674    }
675    return NULL;
676}
677
678
679/*****************************************************************************
680 * Core
681 */
682
683/*
684 * Send a request and process the reply
685 */
686FILE *
687_http_request(struct url *URL, const char *op, struct url_stat *us,
688	      struct url *purl, const char *flags)
689{
690    struct url *url, *new;
691    int chunked, direct, need_auth, noredirect, verbose;
692    int code, fd, i, n;
693    off_t offset, clength, length, size;
694    time_t mtime;
695    const char *p;
696    FILE *f;
697    hdr_t h;
698    char *host;
699#ifdef INET6
700    char hbuf[MAXHOSTNAMELEN + 1];
701#endif
702
703    direct = CHECK_FLAG('d');
704    noredirect = CHECK_FLAG('A');
705    verbose = CHECK_FLAG('v');
706
707    if (direct && purl) {
708	fetchFreeURL(purl);
709	purl = NULL;
710    }
711
712    /* try the provided URL first */
713    url = URL;
714
715    /* if the A flag is set, we only get one try */
716    n = noredirect ? 1 : MAX_REDIRECT;
717    i = 0;
718
719    need_auth = 0;
720    do {
721	new = NULL;
722	chunked = 0;
723	offset = 0;
724	clength = -1;
725	length = -1;
726	size = -1;
727	mtime = 0;
728
729	/* check port */
730	if (!url->port)
731	    url->port = _fetch_default_port(url->scheme);
732
733	/* were we redirected to an FTP URL? */
734	if (purl == NULL && strcmp(url->scheme, SCHEME_FTP) == 0) {
735	    if (strcmp(op, "GET") == 0)
736		return _ftp_request(url, "RETR", us, purl, flags);
737	    else if (strcmp(op, "HEAD") == 0)
738		return _ftp_request(url, "STAT", us, purl, flags);
739	}
740
741	/* connect to server or proxy */
742	if ((fd = _http_connect(url, purl, flags)) == -1)
743	    goto ouch;
744
745	host = url->host;
746#ifdef INET6
747	if (strchr(url->host, ':')) {
748	    snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
749	    host = hbuf;
750	}
751#endif
752
753	/* send request */
754	if (verbose)
755	    _fetch_info("requesting %s://%s:%d%s",
756			url->scheme, host, url->port, url->doc);
757	if (purl) {
758	    _http_cmd(fd, "%s %s://%s:%d%s HTTP/1.1",
759		      op, url->scheme, host, url->port, url->doc);
760	} else {
761	    _http_cmd(fd, "%s %s HTTP/1.1",
762		      op, url->doc);
763	}
764
765	/* virtual host */
766	if (url->port == _fetch_default_port(url->scheme))
767	    _http_cmd(fd, "Host: %s", host);
768	else
769	    _http_cmd(fd, "Host: %s:%d", host, url->port);
770
771	/* proxy authorization */
772	if (purl) {
773	    if (*purl->user || *purl->pwd)
774		_http_basic_auth(fd, "Proxy-Authorization",
775				 purl->user, purl->pwd);
776	    else if ((p = getenv("HTTP_PROXY_AUTH")) != NULL && *p != '\0')
777		_http_authorize(fd, "Proxy-Authorization", p);
778	}
779
780	/* server authorization */
781	if (need_auth || *url->user || *url->pwd) {
782	    if (*url->user || *url->pwd)
783		_http_basic_auth(fd, "Authorization", url->user, url->pwd);
784	    else if ((p = getenv("HTTP_AUTH")) != NULL && *p != '\0')
785		_http_authorize(fd, "Authorization", p);
786	    else if (fetchAuthMethod && fetchAuthMethod(url) == 0) {
787		_http_basic_auth(fd, "Authorization", url->user, url->pwd);
788	    } else {
789		_http_seterr(HTTP_NEED_AUTH);
790		goto ouch;
791	    }
792	}
793
794	/* other headers */
795	if ((p = getenv("HTTP_USER_AGENT")) != NULL && *p != '\0')
796	    _http_cmd(fd, "User-Agent: %s", p);
797	else
798	    _http_cmd(fd, "User-Agent: %s " _LIBFETCH_VER, __progname);
799	if (url->offset)
800	    _http_cmd(fd, "Range: bytes=%lld-", (long long)url->offset);
801	_http_cmd(fd, "Connection: close");
802	_http_cmd(fd, "");
803
804	/* get reply */
805	switch ((code = _http_get_reply(fd))) {
806	case HTTP_OK:
807	case HTTP_PARTIAL:
808	    /* fine */
809	    break;
810	case HTTP_MOVED_PERM:
811	case HTTP_MOVED_TEMP:
812	case HTTP_SEE_OTHER:
813	    /*
814	     * Not so fine, but we still have to read the headers to
815	     * get the new location.
816	     */
817	    break;
818	case HTTP_NEED_AUTH:
819	    if (need_auth) {
820		/*
821		 * We already sent out authorization code, so there's
822		 * nothing more we can do.
823		 */
824		_http_seterr(code);
825		goto ouch;
826	    }
827	    /* try again, but send the password this time */
828	    if (verbose)
829		_fetch_info("server requires authorization");
830	    break;
831	case HTTP_NEED_PROXY_AUTH:
832	    /*
833	     * If we're talking to a proxy, we already sent our proxy
834	     * authorization code, so there's nothing more we can do.
835	     */
836	    _http_seterr(code);
837	    goto ouch;
838	case HTTP_PROTOCOL_ERROR:
839	    /* fall through */
840	case -1:
841	    _fetch_syserr();
842	    goto ouch;
843	default:
844	    _http_seterr(code);
845	    goto ouch;
846	}
847
848	/* get headers */
849	do {
850	    switch ((h = _http_next_header(fd, &p))) {
851	    case hdr_syserror:
852		_fetch_syserr();
853		goto ouch;
854	    case hdr_error:
855		_http_seterr(HTTP_PROTOCOL_ERROR);
856		goto ouch;
857	    case hdr_content_length:
858		_http_parse_length(p, &clength);
859		break;
860	    case hdr_content_range:
861		_http_parse_range(p, &offset, &length, &size);
862		break;
863	    case hdr_last_modified:
864		_http_parse_mtime(p, &mtime);
865		break;
866	    case hdr_location:
867		if (!HTTP_REDIRECT(code))
868		    break;
869		if (new)
870		    free(new);
871		if (verbose)
872		    _fetch_info("%d redirect to %s", code, p);
873		if (*p == '/')
874		    /* absolute path */
875		    new = fetchMakeURL(url->scheme, url->host, url->port, p,
876				       url->user, url->pwd);
877		else
878		    new = fetchParseURL(p);
879		if (new == NULL) {
880		    /* XXX should set an error code */
881		    DEBUG(fprintf(stderr, "failed to parse new URL\n"));
882		    goto ouch;
883		}
884		if (!*new->user && !*new->pwd) {
885		    strcpy(new->user, url->user);
886		    strcpy(new->pwd, url->pwd);
887		}
888		new->offset = url->offset;
889		new->length = url->length;
890		break;
891	    case hdr_transfer_encoding:
892		/* XXX weak test*/
893		chunked = (strcasecmp(p, "chunked") == 0);
894		break;
895	    case hdr_www_authenticate:
896		if (code != HTTP_NEED_AUTH)
897		    break;
898		/* if we were smarter, we'd check the method and realm */
899		break;
900	    case hdr_end:
901		/* fall through */
902	    case hdr_unknown:
903		/* ignore */
904		break;
905	    }
906	} while (h > hdr_end);
907
908	/* we have a hit */
909	if (code == HTTP_OK || code == HTTP_PARTIAL)
910	    break;
911
912	/* we need to provide authentication */
913	if (code == HTTP_NEED_AUTH) {
914	    need_auth = 1;
915	    close(fd);
916	    fd = -1;
917	    continue;
918	}
919
920	/* all other cases: we got a redirect */
921	need_auth = 0;
922	close(fd);
923	fd = -1;
924	if (!new) {
925	    DEBUG(fprintf(stderr, "redirect with no new location\n"));
926	    break;
927	}
928	if (url != URL)
929	    fetchFreeURL(url);
930	url = new;
931    } while (++i < n);
932
933    /* we failed, or ran out of retries */
934    if (fd == -1) {
935	_http_seterr(code);
936	goto ouch;
937    }
938
939    DEBUG(fprintf(stderr, "offset %lld, length %lld,"
940		  " size %lld, clength %lld\n",
941		  (long long)offset, (long long)length,
942		  (long long)size, (long long)clength));
943
944    /* check for inconsistencies */
945    if (clength != -1 && length != -1 && clength != length) {
946	_http_seterr(HTTP_PROTOCOL_ERROR);
947	goto ouch;
948    }
949    if (clength == -1)
950	clength = length;
951    if (clength != -1)
952	length = offset + clength;
953    if (length != -1 && size != -1 && length != size) {
954	_http_seterr(HTTP_PROTOCOL_ERROR);
955	goto ouch;
956    }
957    if (size == -1)
958	size = length;
959
960    /* fill in stats */
961    if (us) {
962	us->size = size;
963	us->atime = us->mtime = mtime;
964    }
965
966    /* too far? */
967    if (offset > URL->offset) {
968	_http_seterr(HTTP_PROTOCOL_ERROR);
969	goto ouch;
970    }
971
972    /* report back real offset and size */
973    URL->offset = offset;
974    URL->length = clength;
975
976    /* wrap it up in a FILE */
977    if ((f = chunked ? _http_funopen(fd) : fdopen(fd, "r")) == NULL) {
978	_fetch_syserr();
979	goto ouch;
980    }
981
982    if (url != URL)
983	fetchFreeURL(url);
984    if (purl)
985	fetchFreeURL(purl);
986
987    return f;
988
989 ouch:
990    if (url != URL)
991	fetchFreeURL(url);
992    if (purl)
993	fetchFreeURL(purl);
994    if (fd != -1)
995	close(fd);
996    return NULL;
997}
998
999
1000/*****************************************************************************
1001 * Entry points
1002 */
1003
1004/*
1005 * Retrieve and stat a file by HTTP
1006 */
1007FILE *
1008fetchXGetHTTP(struct url *URL, struct url_stat *us, const char *flags)
1009{
1010    return _http_request(URL, "GET", us, _http_get_proxy(), flags);
1011}
1012
1013/*
1014 * Retrieve a file by HTTP
1015 */
1016FILE *
1017fetchGetHTTP(struct url *URL, const char *flags)
1018{
1019    return fetchXGetHTTP(URL, NULL, flags);
1020}
1021
1022/*
1023 * Store a file by HTTP
1024 */
1025FILE *
1026fetchPutHTTP(struct url *URL __unused, const char *flags __unused)
1027{
1028    warnx("fetchPutHTTP(): not implemented");
1029    return NULL;
1030}
1031
1032/*
1033 * Get an HTTP document's metadata
1034 */
1035int
1036fetchStatHTTP(struct url *URL, struct url_stat *us, const char *flags)
1037{
1038    FILE *f;
1039
1040    if ((f = _http_request(URL, "HEAD", us, _http_get_proxy(), flags)) == NULL)
1041	return -1;
1042    fclose(f);
1043    return 0;
1044}
1045
1046/*
1047 * List a directory
1048 */
1049struct url_ent *
1050fetchListHTTP(struct url *url __unused, const char *flags __unused)
1051{
1052    warnx("fetchListHTTP(): not implemented");
1053    return NULL;
1054}
1055